Compare AI models

Specifications
Parameters
671B1.6T
Context window
128k1M
API pricing
Input price
—$0.66
Output price
—$1.98
Cached input price
—$0.022
Cheapest input
—$0.0999Relace
Cheapest output
—$1.7424Alibaba
—$203Alibaba
Benchmarks
Humanity's Last Exam · text only, no tools
30.6%42.7%
Overview
CompanyDeepSeekDeepSeek
Release dateDec 1 2025Aug 13 2026
AccessOpen WeightOpen Weight
Model detailsView modelView model

Frequently asked questions

DeepSeek-V4-Pro-0813 leads DeepSeek-V3.2-Speciale on 1 of the 1 benchmark they both report (Humanity's Last Exam). Only DeepSeek-V4-Pro-0813 has a verified first-party price for an available API: $0.66 per million input tokens and $1.98 per million output tokens. No currently available pay-as-you-go API rate is tracked for DeepSeek-V3.2-Speciale. DeepSeek-V3.2-Speciale shipped 255 days before DeepSeek-V4-Pro-0813, so benchmark comparisons should account for the intervening progress.

DeepSeek-V3.2-Speciale has 671B parameters, while DeepSeek-V4-Pro-0813 has 1.6T. Context windows are 128k (DeepSeek-V3.2-Speciale) vs 1M (DeepSeek-V4-Pro-0813).

On Humanity's Last Exam · text only, no tools, DeepSeek-V4-Pro-0813 leads at 42.7% vs DeepSeek-V3.2-Speciale at 30.6%.